A mini book on ORM 2
Everything you need to draw a conceptual schema you can defend — in about an hour of reading.
Object-Role Modeling is not hard, but it is unfamiliar. It asks you to give up the two things most data modelers reach for first — attributes and tables — and to write sentences instead. This book walks through the method in the order you actually need it, and ends with Terry Halpin's seven-step design procedure applied to a real report.
Every diagram here was drawn by the Factum extension itself, and every figure links to the model file behind it. Install the extension, click Open this model, and you can take any figure apart, change a constraint, and watch the verbalization change with it. That loop — draw, read, correct — is the whole method in miniature.
Contents
- 01Why fact-based modelingWhat attributes cost you, and why sentences beat boxes.
- 02Elementary factsThe atom of an ORM model, and how to split a fact that is doing two jobs.
- 03Objects and reference schemesEntity types, value types, reference modes and composite identifiers.
- 04Predicates, roles and readingsArity, role order, and writing readings that verbalize well.
- 05UniquenessThe constraint that carries the most meaning — and the arity check it powers.
- 06Mandatory rolesWhich roles must be played, simple and disjunctive.
- 07The other constraintsValue, frequency, ring, subset, exclusion and equality.
- 08Subtypes and objectificationWhen a subtype earns its place, and when a fact becomes a thing.
- 09The design procedureHalpin's seven steps, worked end to end on a real report.
- 10Pitfalls and a checklistThe mistakes that recur, and a checklist before you map.
How to read it
Chapters 1 to 4 are the vocabulary: facts, objects, roles and readings. Chapters 5 to 8 are the constraints, which is where a model stops being a picture and starts being a specification. Chapter 9 puts it together as a procedure, and chapter 10 is the list of things that go wrong.
If you are impatient, read chapter 2 and chapter 5. Elementary facts and uniqueness constraints carry most of the method between them; everything else is refinement.
What this book is not
It is an introduction, not a reference. ORM has a formal grounding, a textual language (FORML), derivation rules, and a body of transformation theory that a short book cannot cover. For the full treatment, Halpin and Morgan's Information Modeling and Relational Databases is the standard work, and the ORM Foundation is where the community lives.
For how Factum turns a finished model into a database schema, see Mapping rules. For the editor itself, see the guide.